Suppliers of industrial AI agents, data tools and factory software now have 49 state-endorsed reference plants, which local governments are told to help build datasets and agents.
China’s Ministry of Industry and Information Technology (MIIT) and five other central departments published the names of 34 factories on November 19 as the second batch of "pioneer-level" smart factories (领航级智能工厂), the top tier of the country’s four-level smart factory program. Xinhua reported that the new entrants use AI in more than 70% of their production scenarios, have developed more than 1,000 industrial AI agents and together process over 300 million tokens a day. With a first batch of 15 named in 2025, China now counts 49 factories at this level.
The heads of the 34 companies appeared on stage together at the opening of the 2026 World Intelligent Manufacturing Conference in Nanjing, where they launched a joint "pioneer factory cultivation action" (领航工厂培育行动), Science and Technology Daily reported. The list includes Xiaomi Auto, ZTE’s Nanjing plant, Tasly Pharmaceutical, Shougang Jingtang, FAW-Volkswagen, Harbin Electric Machinery and Seres.
The program gives MIIT a set of named, state-endorsed examples of AI adoption on the factory floor. The notice tells local governments to help the selected plants build high-quality datasets, develop industrial AI agents, try out advanced manufacturing methods and create smart supply chains. What is still unknown is how much public money, if any, comes with the designation: the notice, as republished, sets out no funding amount.
Thirty-four plants span aircraft, ships, liquor and apparel
The formal notice, numbered 工信厅联通装函〔2026〕562号 and dated November 17, was issued by the general offices of MIIT, the National Development and Reform Commission, the Ministry of Finance, the State-owned Assets Supervision and Administration Commission, the State Administration for Market Regulation and the National Data Administration. The 21st Century Business Herald republished it from MIIT’s own WeChat channel on November 19. It says the list came out of provincial recommendations, expert review, site checks, on-site defenses and an online public notice. MIIT first published the 34 proposed projects for public comment on October 26, Xinhua reported at the time.
The project names show where AI is already written into the plan. According to the full list in an IT Home report republished by ifeng, Xiaomi Auto’s entry is an "AI-driven agile manufacturing" super smart factory for high-end new energy vehicles. Jiangxi Copper entered an "AI-enabled" copper smelter, Luzhou Laojiao a baijiu plant that combines intangible-heritage brewing skills with AI, and LCFC (Hefei) Electronics an AI PC factory. Heavy industry is heavily represented, with AVIC Xi’an Aircraft, CRRC Sifang’s high-speed train plant, Dongfang Turbine and PetroChina’s Dushanzi petrochemical branch. Shipbuilding is represented by Hudong-Zhonghua, whose entry is a factory for large LNG carriers, and Nantong COSCO KHI.
China News Service said pharmaceutical and apparel makers are on the list for the first time, and that the factories are spread across northeast, northwest, southwest, central and east China. It reported that the second batch raised average labor productivity by 34.4% and production efficiency by 19.3%, and that the 34 plants have pulled nearly 4,000 other companies into digital upgrades.
AI has moved from support work into production itself
Xinhua said AI in these plants has spread from operational support into research and design, process optimization, production operations and production management. The ZTE plant in Nanjing gives the clearest example. Its project is a factory driven by multiple AI agents. According to the Nanjing Morning Post, republished by Sohu, dozens of agents work together there on research, production scheduling, quality inspection and resource allocation. The plant can switch its lines seamlessly between more than 300 product models.
The tier sits on top of a large base. MIIT’s 2026 action notice, published in July, required every applicant to hold excellence-level status already and capped each province’s recommendations at 10% of its cumulative excellence-level projects. Xinhua’s October report said China had more than 13,000 advanced-level and more than 500 excellence-level smart factories at the end of September. Shanghai Securities News, in a report republished by NetEase, put the totals at more than 60,000 basic-level, more than 13,000 advanced-level and more than 750 excellence-level factories.
The push toward agents is consistent with policy MIIT laid out in October, when it said it was drafting measures for new-generation smart manufacturing and would validate factory AI agents task by task, as EastFrontier reported at the time.
Standards bodies and supplier spillover come next
MIIT chief engineer Wang Weiming told the Nanjing conference that the ministry would keep working to deepen smart factory construction, speed up innovation in smart equipment and promote the spread of results, according to a Xinhua Daily report carried by Sohu. At the same event, MIIT’s smart manufacturing standardization technical committee and a Chinese expert committee for IEC international smart manufacturing standards began formation, Science and Technology Daily reported.
The next markers are practical. The 34 plants are sharing technology, solutions and construction experience along their supply chains, the Nanjing Morning Post reported, so purchases of agent platforms, industrial models and data services by their suppliers are where spillover would show. Under the 2026 rules, candidates for the top tier come only from excellence-level factories, within each province’s 10% cap, which makes the growing excellence-level pool the pipeline for future batches. MIIT’s next annual cultivation notice, and whether it raises the bar for AI use above what this batch reports, is the document to watch.